Sunday, May 26, 2024

Reading Ray Kurzweil for a Quarter of a Century: Some Thoughts

I first took an interest in Raymond Kurzweil's writing about the Singularity back when he first published The Age of Spiritual Machines (1999). All that sort of thing was a lot more novel then, and I was really intrigued by it--the more in as Kurzweil, unlike so many futurologists, presented long lists of specific forecasts precisely dated for 2009, 2019 and after.

Those forecasts for dates that were not too far away in particular had me thinking "Let's see what happens with all this," and so as I took in news about technological developments kept an eye open for signs that the advances he described--in the neural net-based pattern recognition that seemed most central to the progress in artificial intelligence, the mass-production and mass-scale usage of carbon nanotubes that would power Moore's Law into a post-Silicon age, the construction of nano-scale machinery, etc.--were actually happening.

A few years on, I saw no evidence of anything remotely like the pace of progress he predicted. The Segway scooter, instead, was as good as that seemed to get. Kurzweil's The Singularity is Near (2005), which seemed to me to elaborate his earlier theorizing rather than really break new ground, let alone demonstrate dramatic progress since the time of his last book, did not convince me otherwise. Nor did anything else I saw in the years that followed, including the smart phones that, for all their wonders, refined well-established technologies rather than representing the breakthroughs he talked about. And indeed, when 2009 rolled around, ten years after the appearance of Spiritual Machines, I published a systematic evaluation of his technological forecasts for that year in which I concluded that he had consistently been far off the mark in those areas where it counted most.

Indeed, watching so long and seeing so little I was getting fairly skeptical of promises for radical developments just around the corner. In fact I paid a lot less attention to technological forecasting afterward--in part because the writing about it I saw gave me less reason to do so, the writers seemingly saying the same things over and over again rather than presenting any new ideas, or new arguments for them, let alone present evidence for momentous things happening.

Of course, around the mid-2010s there was a new wave of techno-hype, the more in as progress in neural nets and machine learning was gaining steam, all as we were told a great wave of automation was very likely to soon sweep through the economy--with the supposedly imminent automation of our vehicles merely the beginning. At its height I suggested, optimistically, that perhaps we were running a decade or so behind Kurzweil's forecasts.

Alas, the it was just another bubble of techno-hype that went bust with almost nothing to show for it before the end of the decade, as the pandemic reminded us all how un-automated our economy really was, and seemed likely to remain for a good long while to come, while we ran a lot more than a decade behind Kurzweil's forecasts--our 2019 less advanced than his 2009, giving us additional grounds to doubt that it was worth comparing our trajectory with his timelines at all.

Of course, I knew that techno-hype is a cyclical thing and would resurge again, but it happened sooner than I would have guessed, while frankly being prompted by rather less than I thought it would take to accomplish that--Open AI and its chatbots. That excitement remains very much with us, fed by further advances in chatbots, and promised wonders such as what we have seen of Sora, while the price of the NVIDIA corporation's stock soared past the $1,000 mark because of its standing as a maker of AI chips, all as Wall Street generally seems to be behaving not as if the Singularity is near, but as if the Singularity is here (with the apocalyptic panic coming from some quarters only the other side of the expectation that momentous things were happening). The result is an especially opportune climate for Kurzweil's long-promised and repeatedly delayed follow-up to The Singularity is Near--The Singularity is Nearer--to appear before the world. As yet only the earliest reviews have appeared, as those of us who did not get early copies must wait another month before having the chance to check out Kurzweil's book for ourselves.

Will Kurzweil persuade those of us who grew skeptical of him over this uninspiring past quarter of a century to pay attention to his predictions again?

I suppose we will find that out soon enough.

No comments:

Subscribe Now: Feed Icon